$74 NOTES VOL. 26

EXPERIMENTAL^^ S-EthylenedWxy6-androsten-17-one (I). To a chilled solu-

tion of 1.64 g. (4.94 mmolea) of 3-ethylenedioxy-17fl-hy- droxy-5-androstene'd in 15 cc. of dry pyridine waa added the chromic oxide-pyridine reagents prepared from 2.06 g. of chromic oxide and 15 cc. of pyridine. The stirred mixture was kept a t room temperature over night and wm then added to 100 cc. of ice water. Chloroform was added and the mixture was filtered through diatomaceous earth. The pre- cipitate waa washed with more chloroform and the layers of the filtrate were separated. The chloroform phase was washed with water, and waa dried and partially decolorized over magnesium sulfate and activated charcoal. The solution waa filtered and freed from solvents by evaporation. The residue was crystallized from methanol to afford 1.18 g. (73y0), m.p. 189-197' (lit. m.p. 185-192',4c 197-198°4d).

16-Ethoxalyl-5-ethylenedioxy-5-and~osten-17-one (11). Sodium (75 mg.) was dissolved with heating in a mixture of 30 cc. of benzene and 2 cc. of methanol. Solvent was distilled from the stirred mixture until the distillation temperature reached 80'. To the resulting stirred suspension, cooled to room temperature, was added 1 g. (3 mmoles) of &ethylene- diosy-5androsten-17-one (I) and 1 cc. of ethyl oxalate. The mixture was stirred at room temperature for 16 hr. and was then estracted with several portions of 1 % aqueous potassium hydroxide solution. The extracts were neutralized with 30% aqueous sodium dihydrogen phosphate solution and the mixture was extracted several times with chloro- form. The combined chloroform extracts were washed with water, dried, and evaporatcd. The residue was crystallized from ether to afford 0.795 g. (72%) of white solid with m.p. 161-16-1". A sample recrystallized from ether showed m.p. 161-163; [a]: -103' ( c , 0.793 in chloroform); Xmsx 5.76 p (ester), 5.04 p, and 6.17 p (enoliaed 8-diketone system); posi- tive enol test.

Anal. Calcd. for C25H3106: C, 60.74: H, 7.96. Found: C, 69.08; H, 8.20. 3-Ethvlenediori1-l6~-Ruoro-5-androoten-l~-one (111). To a

suspension of 860mg: [2 mmoles) of 16-ethoxalyl:3-ethylene- dioxy-5-androsten-17-onc (IT) in 20 cc. of methanol, cooled to - 15", was added 3 cc. of a IN mcthanolic sodium meth- oxide solution and the stirred, cooled solution was saturated with a rapid stream of perchloryl fluoride gas.11 Nitrogen was then blown through the solution to remove excess per- chloryl fluoride and the reaction mixture (neutral, negative enol test) was evaporated in Vacuo at room temperature. The residue was dissolved in chloroform and water, and the layers were separated. The organic phase was dried and evaporated and the residue waa redissolved in 20 cc. of methanol. Potassium acetate (1.5 g.) was added and the stirred mixture was heated under reflux for 1 hr. The solvent was evaporated and the residue was dissolved in chloroform and water. The chloroform layer was washed with a little water and was dried and evaporated to afford 913 mg. of a glass. This was crystallized from 10 cc. of ether and the solid was collected and washed twice with 5-cc. portions of ether. There was obtained 170 mg. (24%) with m.p. 227-230'. The analytical sample was recrystallized twice from methyl- ene chloride-ether; m.p. 240-243O; [a]:: + 26.8" (c , 1.64 in chloroform); Xmax 5.70 p (17-one), 9.08 p (3-ketal).

(10) Melting points wcrc taken 011 a IZoHcr micro hot- dage and are corrected. Ultraviolet spectra wcre deter- mined in methanol on a Cary recording spectrophotometer :knd infrared spcctra (potassium bromide disks) on a Perkin Elmer spectrophotometer (Model 21). An ethanolic solution of ferric chloride was used for the enol test. Solutions were dried over magnesium sulfate and evaporated under reduced pressure.

(11) V. Papesh, Chem. Eng. News, 37, (NO. 28), GO (1959) hits reported an explosion resulting from the addition of sodium methoxide to a vessel containing the mixed vapors of methanol and percliloryl fluoride.

Anal. Calcd. for C2tH2pFOa: e, 72.37; H, 8.39; F, 5.45. Found: C, 72.01; H, 8.73; F, 5.82.

Total evaporation of the ether mother liquors and tri- turation with a small amount of fresh ether afford 360 mg. of a solid with m.p. 100-105'. This was recrystallized from acetone-hexane and from ether; m.p. 109-111"; [a]? -63.5' (c , 0.38 in chloroform); Xmax 2.83 p (OH), 5.70 p (shoulder, a-fluoroketone), 5.78 L( (ester?), 7.90-8.02 p . , . . .. (ester?), 9.08 p (3-ketal).

Anal. Calcd. for CzlH2pF05.H20: C, 63.29; H, 7.84; F, 4.77; H20, 4.52. Found: C, 62.86; H, 7.87; F, 4.53; HzO, 4.84.

1 6~-Flzioro-4-androslenc-SJ17-dione ( 16~-Flzio~oleslosterone, IV). A solution of 348 mg. (1 mmole) of 3-ethylenedioxy- lG~-fluoro-5-androsten-l7-0ne (111), in 10 cc. of methanol and 0.5 cc. of water was reduced with 175 mg. of sodium borohydride at the reflux point for 3 hr. The mixture was cooled and poured into 25 cc. of water to give a suspension which was extracted with four 10-cc. portions of chloroform. The combined extracts were washed with water, dried and evaporated. The residue was crystallized once from methanol to afford 0.3 g. (E%%) with m.p. 189-195'; A,,, 2.80 p (OH region), no absorption in the C = O region. The solid was redissolved in 20 cc. of methanol containing 1 cc. of 8% aqueous sulfuric acid and the mixture was allowed to reflux for 2 hr. and was then diluted with an additional 30 cc. of methanol. Duolite A-4 anion exchange resin (OH form)'* was added with stirring until the solution was neutral. The resin W R ~ removed by filtration and was washed well with methanol. The combined filtrate and washings were evapo- rated and the residue was dissolved in chloroform and water. The chloroform phase was washed with a little water and was dried and evaporated to give a crystalline residue which was recrystallized from ether to afford 0.17 g. ( 5 5 w over-all from 111), m.p. 154-156"; [a]? $117" (c, 0.5in chloroform) (lit. valuesa 'for 16a-fluorotestosterone: m.p. 153-158', [CY]D +113' in chloroform); Xmsx 240 mp (e, 15,800); X,,, 2.94 p (OH), 6.01 p (A4-3-one).

Anal. Calcd. for C1sHt,O?F: C, 74.47; H, 8.88; F, 6.20. Found: C, 74.46; H, 9.03; F, 6.00.

Acknowledgmcnt. We would like to thank Mr. W. Fulmor and staff for spectrophotometric and polarimetric data and Mr. L. Brancone and staff for microanalyses.

ORGANIC CHEMICAL RESEARCH SECTION LEDERLE LABORATORIES DIVISION AMERICAN CYANAMID Co. PEARL RIVER, N. Y.

(12) Duolite A-4 is the trademark of the Chemical Process Co:, Redwood City, Calif., for a weakly basic anion exchange resin.

6@-Hydroxylation of 9a-Fluorohydrocortieone

LELAND L. SMITU,~ JOSEPH J. GOODMAN, HAROLD MENDELBOHN, JOHN P. DUSZA, AND SEYMOUR BERNBTEIN

Received June 16,1060

In a study of the hydroxylation of Da-fluorohydro- cortisone (I) by LL variety of Streptomyces species we have found that a strain of S. rimosus (Lederle Laboratories Collection No. T1686B) affords as a major product a very polar reducing monohy- droxylated 9a-fluorohydrocortisone I1 different from the previously described lt- and l6a-hy-

(1) Prevent address: Wyeth Laboratories, Inc., Philadel- phia, Pa.

NOTES 975

drosy-Da-fluorohydrocortisones."8'2b Isolation of the polar steroid I1 was accomplished via partition chromatography. In addition a more polar reduc- ing steroid was recovered from later column fractions.

The polar steroid If was recognized as retaining the dihydroxyacetone sidechain by comparison of its reducing capacity towards alkaline Tetra- zolium Blue over a period of several hours versus that of 9a-fluorohydrocortisone (essentially iden- tical time curves) and by consideration that I1 gave a typical Porter-Silber chromogen. A hypso- chromic shift of 7 nip in the ultraviolet absorption spectra of I1 indicated the point of attachment of the additional hydroxyl group was involved in the A-ring chromophore, the GP-position being suggested. Confirmation of the @-hydroxy form- ulation was made by a study of ultraviolet light absorption spectra in alkaline e tha~io l ,~ wherein spectra characteristic of 6p-hydroxy-A4- 3-ketosteroids were obtained.

The polarity of the monohydroxy-9a-fluoro- hydrocortisone I1 in several paper partition chrom- atographic systems was strikingly more polar than the known 15-hydroxy- and 16a-hydroxy-9a- fluorohydrocortisoiies, which is consistent with the suggested Gp-hydroxy structure for 11. The structure of I1 is thus assigned that of Sa-fluoro- Gp-hydroxyhydrocortisoiie (Ya-fluoro-6~,11/3,17a,- 31-tetrahydroxy-4-pregnene-3,2O-dione).

.An opportunity for a complete proof of struc- turc of the Gp-hydroxy derivative 11, a t the same time offering a rionfermciitative preparative method for the compound from 9a-fluorohydrocortisone 21-acetate (Ia) followed from our finding that the &methyl enol ether I11 (not isolated) of Ia was oxidized by monoperphthalic acid directly to the ($-hydroxy-A4-3-ketosteroid IV. This appears to he the first report of the reaction of an enol ether with a peracid t o yield the respective 6P-hydroxy- A4-3-ketonc. Similar reaction of a Aat5-3-enol zcetnte to yield the (jp-hydroxy-A4-3-ketonc system has been rccorded previously.

Saponification of the GF-hydroxy-21-acetate It' afforded Da-fluoro-G/3-hydroxyhydracortisone (11) identical with the product 11 isolated from S. riniosus hydroxylutioii of Da-fluorohydrocortisone."

(%a) W. S. hlcAlccr, !c1. .I. Koalownki, T. II. Stoiidt, tliitl J . hI. Clicnicrtla, J. Ory. Chciih., 23, 508 (1!)58).

(211) S. lhiiqteiri, 11. 11. Letihrd, W. S. Allcii, hI. H d c r , l<. Littell, ;J. \I. Stelar, L. I. E'c~ldmtii, mid It. I f . I h i k , J . 1 t / L , C'/w/ ih. Soc., 78, 5693 ( J !)56) ; 81, l(i8'J ( I XI!)).

(8) A. S. Mcyer, J . Org. Chern., 20, 1240 (1955). (4) J. Ilomo, G. Roscnkrantx, C. Djcrassi, and F. Sorid-

licimcr, J. Orq. Cheni., 19, 1500 (1954). (5) Gp-IIydroxy-Dru-fluorohydrocortisonc has been iso-

lated (without ch:tracterization or proof of structurc) by McAlcer ct ~ 1 . , * ~ from 8treploniyces s p . fcrmentutioiis with Ya-fiuorohydrocortisorlc and by N. E. Rigler, A. J. Shay, L. I. Feldmari, and 13. Niclscii of these laboratorics from Xortkrelkz s p . fcrmciitntioils with the same substrate.

C H a R I co

I I...ALl HO I I

CHzC H I co

L

I11

CH~OAC I

Monoperph-

arid ~

OH IV

EXPERIMENTAL

Fermentation of Sa-flwrohvdrowrtisone. An inoculum medium consisting of coiu steep liquor (20 g./l.), sucrose (30 g./l.), calcium carbonate (7 g./l.), and ammonium sul- fate (2 g./l.) was prepared and inoculated with spores of 8. rimosus (Lederle Laboratories Collection No. T1686B). After 24 hr. of growth the inoculum was used at 401, levels to inoculate medium consisting of corn steep liquor (25 g./l,), starch (40 g./l.), calcium carbonate (5 g./l.), and lard oil (0.2'76 by volume). After a further 24 hr. of growth a solu- tion of Sa-fluorohydrocortisone dissolved in propylene glycol was added in such an amount that the final steroid concentra- tion was 500 pg./ml. After 48 hr. of aerated fermentation, analyses indicated that cu. 430 pg./ml. of steroid p r d u c t had been formed, and after 72 hr., 375 pg./ml. Paper chroma- tographic analyses indicated that the major product wm the polar 6p-hydroxy-9a-fluorohydrocortisone I1 (R, 0.25 in System 116) together with a lesser amount of a still more polar reducing steroid (RI 0.06). No unaltered I was de- tected.

The hariested broth (fermentation time 48 hr., 1330 ml.) was filtered through diatomaceous earth, the filter cake wmhed with water, methanol, and chloroform, and the combined filtrates extracted with equal volumes of ethyl acetate three times. The combined extracts were evaporated under vacuum to a viscous residue which analyzed on paper chromatograms as a mixture of about ten different steroidal components, of which the major component wm the 6phy- droxy derivative 11. The total steroid sample was partitioned on Celite diatomaceous earth using the system dioxane-cy- clohexancnster (5: 2: l).B Two major reducing products were eluted, the desired 6p-hydruxy derivative I1 at hold- back voluntc 2.1, and a more polar reducing steroid present in lrssw :mounts at hold-back volume 3.2. The appropriate fractions were evaporated uritier vacuum to yield 47 mg. of cywtdIirie 6p-hydroxg-0a-fluorohydrocortisone (Il), m.p. 2O(i-20!io. Aftm several rccrystalliwttions from methanol/ acetone tlic product nrcltesd at 222-224' dec. (Kofler block), 224-227' dec. (capillary); [a] + 56" (methanol); Amax 232 mp (in ethanol); A'.;;'' (E;:$,) at 2 hrs., 259 mp (239), 337 mp (317), 405 mp (203), 442 mp (200, infleotion), 470 mp (210), 500 mg (177, inflection); A::: 2.90 p , 5.84 p, 6.00 p, 6.12 p, 9.47 p, 10.04 p, 10.G2 p, 11.30 p, ctc.

(6 ) L. L. Smith, T. Focll, 11. De Maio, atid M. Halwer, J . Am. Pharni. Assoc., 48, 528 (1059).

VOL. 26 970 NOTES

Papcrgram mobilities0 relative to 16a-hydroxy-Ya-fluoro- hydrocortisone include: Systcm I, 0.35; System 11, 0.48; System 111, 0.42. The steroid reduces alkaline Tetrafiolium Blue reagent and gives a yellow fluorescent reaction with isonieotinir acid hydrazide' on papergrams. A normal Porter- Silber chromogen* with A,, 413 mp (E;L. 395) us. A,, 414 mp. (E:.",. 374) for Sefluorohydrocortisone is found. Maxi- mum color development with alkaline Tetrazolium Blues occurred at 30 min., with Amax 520 mp (E:E. 580) us. A,,, 525 mp 654) for 9a-fluorohydrocortisoner also maxi- mum at 30 min.

Spectra in 0.066N ethanolic potassium hydroxide at room temperature (22') showed a slow shift. from Amax 233 mp (EiZ. 312) at 3 min. to A,. 235 mp @ik. 226), 380 mp (28) at 24 hr. At 60' the spectra changed from Amax 232 mp (E;:. 308) at 3 min., A,. 236 mp (172) a t 1 hr., Amax 239 mp (132)) 378 mp (62) a t 2 hr., to A,, 248 mp (142)) 378 mp (91) a t 3 hr., typical of spectpal changes associated with the 68- hydroxy-A4-3-ketone chromophore.#

The fractions cluted about hold-back volume 3.2 were re- duced in volume under vacuum, analyzing by paper chroma- tography as a single major reducing component of R, 0.06 in System 11. Crystallization of the material from acetone- chloroform-ethyl acetate (1: 1:2) yielded 7.1 mg. of crystale, whose infrared absorption Rpectra indicated the same general structural features to be present as were present in the npec- tra of 11. No further characterization was made of this com- ponent.

Il-Acetol y-9a-$uoro-6&11& 17a,81-tetruhwdrox~-4-preg- nened,80dione ZV. Ten grams of Sa-fluorohydrocortisone 21-acetate (Is) was added to a solution of 75 ml. of dioxane, 10 ml. of trimethyl orthoformate, and 0.5 ml. of absolute methanol.'To this suspension waa added 10 ml. of dioxane containing 0.5 ml. of concd. sulfuric acid. After 10 min. a clear solution resulted. After 20 min. at room temperature pyridine was added dropwiso until the deep red color of the solution waa discharged (total pyridine used, about 1 ml.). The reaction mixture waa poured into water, giving an oily material which waa extracted into ether the ether extract washed- with saline, and then dried. The ether volume was increased to 200 ml. and 120 ml. of 0.31N monoperphthalic acid in ethor waa added. The reaction mixture was then stored in the dark for 15 hr., a t which time the precipitated crystalline product was filtered and washed with ether (4.4 g. weight). Recrystallization from ethyl acetate yielded 1 635 g. of pure IV, m.p. 248-252' dec. (capillary); [a] : + 70' (pyridine); A,, 232 mp (e 14,000); 2.87, 5.72 (Sh), 5.87, 5.92 (Sh), 5.97, 8.06, 10.04, and 10.61 p.

Anal. Calcd. for C2aHs107F: C, 63.00; H, 7.13; F, 4.33. Found: C, 62.46; H, 7.26; F. 4.44.

The mother liquors on standing depositcd an additional 0.60 g. of crude product, which wm recrystallized from ethyl acetatelheptane, 0.21 g., m.p. 234-239' (capillary).

.9a-Fluoro-6&11 @ , I 7a,dl-tetrahy~rox1/-4- pregnene-J,dO-dione (11). A solution of 600 mg. of the 21-acetate IV in 100 ml. of absolute methanol waa prcpared undcr a nitrogen at- mosphere. To this solution was added 3.0 ml. of a 10% aque- ous potamium carbonate solution. After 1 hr. (undcr nitro- gen) the reaction mixture was neutralized with acrtic acid, most of the methanol removed under vacuum, and the reac- tion concentrntc was poured into water. The solids rccovcred by filtration were recrystallized scveral times from ethyl acetateheptanc, yielding 145 mg. of platcs, m.p. 214-244" (capillary), which after drying in vacuum melted a t 235-2399 dec. (capillary), 223-225' dec. (Kofler block); [a]? + 57.7' (pyridine); Amax 232 mp (e 15,000).

(7) L. L. Siiiith and T. Pocll, ATLUZ. Chem., 31, 109

( 8 ) C. C. I'ortcr arid It. H. Silbcr, J . ,!.?id. C / 6 C f / l . , 185,

(1)) L. L. Smith arid M. I I d w r , J . A M . I%u/ t t~ . Awe.,

(1051)).

201 (1950).

48, 348 (1959).

Anal. Calcd. for C21H&F: C, 63.62; H, 7.37; F, 4.79. Found: C, 63.54; H, 7.60; F, 4.65.

The idcntity of the microbiologically derivcd I1 with I1 dcrived chemically was established by comparison of infrared spcctra, sulfuric acid spectra, melting points and papergram mobility, and color test behavior.

CHEMICAL PROCESS IMPROVEMENT DEPARTMENT AND THE

LEDERLE LABORATORIES AMERICAN CYANAJHID Co. PEARL RIVER, N. Y.

ORGANIC CHEMICAL XISEARCR SECTION

Chloromethylation. A Novel Route to &Methyls teroids

JOHN 11. FRIED, ANTHONY N. NUTILE. AND GLEN E. AHTII

Received June 87, 1960

Enol acetates of A4-3-ketosteroids have been utilized to introduce halogen,*12 nitro and hydroxyl substituents at c6 of steroids by an electrophilic process. An extension of this reaction to include carbonium ions could be a useful path to the bio- logically important C6 alkylated steroids.

The solvolytic reaction of chloromethyl methyl ether in acetic acid appeared to be a suitable source of an electrophilic fragment, the chloromethyl car- bonium A solution of this reagent and 17a,- 20,20,2 1-bisme thylenedioxy-3-ace toxy- 3, 5 - pregna- diene-1 1 -one, 11, afforded a crystalline product after percolation through alumina. This product was shown to be a mixture by NMR4 analysis but could not be resolved by recrystallization or chroma- tography. However, reduction of the crude crystal- line product with zinc in acetic acid afforded 4- methylcortisone BMD,s V, in ca. 11% overall yield. It follows that 4-chloromethylcortisone BhllD, 111, was one of the components of the chlorometh- ylation mixture. All attempts to isolate this product failed. However, repeated crystallizations from methanol did afford a second component of the mixture IV in addition to hydrogen chloride. A simpler procedure for isolating IV was to reflux the mixture with dilute hydrochloric acid in mcth- anol. This treatment destroyed the 4-chloromethyl-

(1) A. Bowers, L. C. Ibancs, and H. J. Ringold, J . A VL. Chem. SOC., 81, 3707 (1O5O) and references cited.

(2) B. hf. Bloom, V. V. Bogert, and R. Pinson, Cheiri. Ind., 1317 (1950).

(3) Thc attacking spwics is rcprescnted ns +cIj~C;l as :I mattcr of convenience but could cqually well be CII,O +C:II,Cl

or one of several ion pairs. (4) NMR spectra were run on a V:trian 60MC Spcctroni-

cter a t a concentration of ca. 20 mg. in 0.3 ml. dctitero- chloroform. T = 7/60 + 3.5 whcrc 7 is the observcd band position in c.p.5. rclative to benzene as cxtcrnal staiitJ:ird. Cf. G. V. D. Tiers, J . Phys. Chem., 62 , 1151 (1058). M'c \cisIi to th:tnk I3. Arison mid Dr. N. I t . l'rerincr for the t1ctt.r- mination and intcrprctatioti of tlic NhlR spectra.

(5) N. G. Stc4nlwg, R. 1Iirwhiiwn, and J. M. Chcvncrcl:~, Chern. I nd., 975 (1958).

€I